def __init__(self, show_img=True):
self.pose_estimator = PoseEstimator()
self.object_detector = ObjectDetectionYolo()
self.img = []
self.img_black = []
self.show_img = show_img
self.BBV = BBoxVisualizer()
def __init__(self, show_img=True):
self.pose_estimator = PoseEstimator()
self.object_detector = ObjectDetectionYolo()
self.object_tracker = ObjectTracker()
self.BBV = BBoxVisualizer()
self.KPV = KeyPointVisualizer()
self.IDV = IDVisualizer(with_bbox=False)
self.img = []
self.img_black = []
self.show_img = show_img
def __init__(self, show_img=True):
self.object_detector = ObjectDetectionYolo(cfg=opt.yolo_cfg, weight=opt.yolo_weight)
self.object_tracker = ObjectTracker()
self.BBV = BBoxVisualizer()
self.IDV = IDVisualizer(with_bbox=False)
self.boxes = tensor([])
self.boxes_scores = tensor([])
self.img_black = np.array([])
self.frame = np.array([])
self.id2bbox = {}
self.show_img = show_img
self.CNN_model = CNNInference()
def __init__(self, show_img=True):
self.object_detector = ObjectDetectionYolo(cfg=yolo_cfg, weight=yolo_weight)
self.object_tracker = ObjectTracker()
self.pose_estimator = PoseEstimator(pose_cfg=pose_cfg, pose_weight=pose_weight)
self.BBV = BBoxVisualizer()
self.KPV = KeyPointVisualizer()
self.IDV = IDVisualizer(with_bbox=False)
self.boxes = tensor([])
self.boxes_scores = tensor([])
self.img_black = np.array([])
self.frame = np.array([])
self.id2bbox = {}
self.kps = {}
self.kps_score = {}
self.show_img = show_img
def __init__(self, resize_size, show_img=True):
self.object_detector = ObjectDetectionYolo(cfg=config.yolo_cfg, weight=config.yolo_weight)
self.object_tracker = ObjectTracker()
self.pose_estimator = PoseEstimator(pose_cfg=config.pose_cfg, pose_weight=config.pose_weight)
self.BBV = BBoxVisualizer()
self.KPV = KeyPointVisualizer()
self.IDV = IDVisualizer()
self.boxes = tensor([])
self.boxes_scores = tensor([])
self.frame = np.array([])
self.id2bbox = {}
self.CNN_model = CNNInference()
self.kps = {}
self.kps_score = {}
self.show_img = show_img
self.resize_size = resize_size
def __init__(self, path=config.video_path, show_img=True):
self.pose_estimator = PoseEstimator()
self.object_detector = ObjectDetectionYolo()
self.BBV = BBoxVisualizer()
self.IDV = IDVisualizer()
self.object_tracker = ObjectTracker()
self.video_path = path
self.cap = cv2.VideoCapture(self.video_path)
self.img = []
self.img_black = []
self.show_img = show_img
self.locator = Locator([1, 2])
class ImgProcessorForbbox:
def __init__(self):
self.object_detector = ObjectDetectionYolo()
self.img = []
self.img_black = []
self.BBV = BBoxVisualizer()
def process_img(self, frame, enhanced=None):
with torch.no_grad():
inps, orig_img, boxes, scores, pt1, pt2 = self.object_detector.process(
frame)
try:
img = self.BBV.visualize(boxes, frame)
return img
except:
return frame
class ImgProcessor:
def __init__(self, show_img=True):
self.pose_estimator = PoseEstimator()
self.object_detector = ObjectDetectionYolo()
self.img = []
self.img_black = []
self.show_img = show_img
self.BBV = BBoxVisualizer()
def process_img(self, frame):
with torch.no_grad():
inps, orig_img, boxes, scores, pt1, pt2 = self.object_detector.process(
frame)
if boxes is not None:
key_points, img, img_black = self.pose_estimator.process_img(
inps, orig_img, boxes, scores, pt1, pt2)
img = self.BBV.visualize(boxes, img)
return key_points, img, img_black
else:
return [], frame, cv2.imread("video/black.jpg")
class ImgProcessor:
def __init__(self, show_img=True):
self.pose_estimator = PoseEstimator()
self.object_detector = ObjectDetectionYolo()
self.object_tracker = ObjectTracker()
self.BBV = BBoxVisualizer()
self.KPV = KeyPointVisualizer()
self.IDV = IDVisualizer(with_bbox=False)
self.img = []
self.img_black = []
self.show_img = show_img
def init_sort(self):
self.object_tracker.init_tracker()
def __process_kp(self, kps, idx):
new_kp = []
for bdp in range(len(kps)):
for coord in range(2):
new_kp.append(kps[bdp][coord])
return {idx: new_kp}
def process_img(self, frame):
# frame = cv2.resize(frame, config.frame_size)
img = copy.deepcopy(frame)
img_black = cv2.imread('video/black.jpg')
with torch.no_grad():
inps, orig_img, boxes, scores, pt1, pt2 = self.object_detector.process(
frame)
if boxes is not None:
key_points, img, black_img = self.pose_estimator.process_img(
inps, orig_img, boxes, scores, pt1, pt2)
if config.plot_bbox:
img = self.BBV.visualize(boxes, frame)
if key_points is not []:
id2ske, id2bbox = self.object_tracker.track(
boxes, key_points)
if config.plot_id:
img = self.IDV.plot_bbox_id(id2bbox,
copy.deepcopy(img))
# img = self.IDV.plot_skeleton_id(id2ske, copy.deepcopy(img))
if config.track_idx != "all":
try:
kps = self.__process_kp(id2ske[config.track_idx],
config.track_idx)
except KeyError:
kps = {}
else:
kps = id2ske
#
# if config.plot_kps:
# vis_kps = self.KPV.dict2ls(kps)
# img = self.KPV.vis_ske(orig_img, vis_kps, kp_score)
# img_black = self.KPV.vis_ske_black(orig_img, vis_kps, kp_score)
return kps, img, img_black
else:
return {}, img, img_black
else:
return {}, frame, frame
def __init__(self):
self.object_detector = ObjectDetectionYolo()
self.img = []
self.img_black = []
self.BBV = BBoxVisualizer()
class HumanDetection:
def __init__(self, show_img=True):
self.object_detector = ObjectDetectionYolo(cfg=opt.yolo_cfg, weight=opt.yolo_weight)
self.object_tracker = ObjectTracker()
self.BBV = BBoxVisualizer()
self.IDV = IDVisualizer(with_bbox=False)
self.boxes = tensor([])
self.boxes_scores = tensor([])
self.img_black = np.array([])
self.frame = np.array([])
self.id2bbox = {}
self.show_img = show_img
self.CNN_model = CNNInference()
def init_sort(self):
self.object_tracker.init_tracker()
def clear_res(self):
self.boxes = tensor([])
self.boxes_scores = tensor([])
self.img_black = np.array([])
self.frame = np.array([])
self.id2bbox = {}
def visualize(self):
self.img_black = cv2.imread('video/black.jpg')
if config.plot_bbox and self.boxes is not None:
self.frame = self.BBV.visualize(self.boxes, self.frame, self.boxes_scores)
# cv2.imshow("cropped", (torch_to_im(inps[0]) * 255))
if config.plot_id and self.id2bbox is not None:
self.frame = self.IDV.plot_bbox_id(self.id2bbox, self.frame)
# frame = self.IDV.plot_skeleton_id(id2ske, copy.deepcopy(img))
return self.frame, self.img_black
def process_img(self, frame, gray=False):
self.clear_res()
self.frame = frame
with torch.no_grad():
if gray:
gray_img = gray3D(copy.deepcopy(frame))
box_res = self.object_detector.process(gray_img)
else:
box_res = self.object_detector.process(frame)
self.boxes, self.boxes_scores = self.object_detector.cut_box_score(box_res)
if box_res is not None:
self.id2bbox = self.object_tracker.track(box_res)
return self.id2bbox
def classify_whole(self):
out = self.CNN_model.predict(self.img_black)
idx = out[0].tolist().index(max(out[0].tolist()))
pred = opt.CNN_class[idx]
print("The prediction is {}".format(pred))
def classify(self, frame):
for box in self.id2bbox.values():
x1, y1, x2, y2 = int(box[0]), int(box[1]), int(box[2]), int(box[3])
x1 = 0 if x1 < 0 else x1
y1 = 0 if y1 < 0 else y1
x2 = frame.shape[1] if x2 > frame.shape[1] else x2
y2 = frame.shape[0] if y2 > frame.shape[0] else y2
img = np.asarray(frame[y1:y2, x1:x2])
# cv2.imshow("cut", img)
# cv2.imwrite("img/tmp/0.jpg", img)
out = self.CNN_model.predict(img)
idx = out[0].tolist().index(max(out[0].tolist()))
pred = opt.CNN_class[idx]
print(pred)
text_location = (int((box[0]+box[2])/2)), int((box[1])+50)
cv2.putText(frame, pred, text_location, cv2.FONT_HERSHEY_SIMPLEX, 1, (100, 100, 255), 2)
class HumanDetection:
def __init__(self, resize_size, show_img=True):
self.object_detector = ObjectDetectionYolo(cfg=config.yolo_cfg, weight=config.yolo_weight)
self.object_tracker = ObjectTracker()
self.pose_estimator = PoseEstimator(pose_cfg=config.pose_cfg, pose_weight=config.pose_weight)
self.BBV = BBoxVisualizer()
self.KPV = KeyPointVisualizer()
self.IDV = IDVisualizer()
self.boxes = tensor([])
self.boxes_scores = tensor([])
self.frame = np.array([])
self.id2bbox = {}
self.CNN_model = CNNInference()
self.kps = {}
self.kps_score = {}
self.show_img = show_img
self.resize_size = resize_size
def init(self):
self.object_tracker.init_tracker()
def clear_res(self):
self.boxes = tensor([])
self.boxes_scores = tensor([])
self.frame = np.array([])
self.id2bbox = {}
self.kps = {}
self.kps_score = {}
def visualize(self):
img_black = np.full((self.resize_size[1], self.resize_size[0], 3), 0).astype(np.uint8)
if config.plot_bbox and self.boxes is not None:
self.BBV.visualize(self.boxes, self.frame)
if config.plot_kps and self.kps is not []:
self.KPV.vis_ske(self.frame, self.kps, self.kps_score)
self.KPV.vis_ske_black(img_black, self.kps, self.kps_score)
if config.plot_id and self.id2bbox is not None:
self.IDV.plot_bbox_id(self.id2bbox, self.frame)
self.IDV.plot_skeleton_id(self.kps, self.frame)
return self.frame, img_black
def process_img(self, frame, gray=False):
self.clear_res()
self.frame = frame
with torch.no_grad():
if gray:
gray_img = gray3D(copy.deepcopy(frame))
box_res = self.object_detector.process(gray_img)
else:
box_res = self.object_detector.process(frame)
self.boxes, self.boxes_scores = self.object_detector.cut_box_score(box_res)
if box_res is not None:
self.id2bbox = self.object_tracker.track(box_res)
self.id2bbox = eliminate_nan(self.id2bbox)
boxes = self.object_tracker.id_and_box(self.id2bbox)
inps, pt1, pt2 = crop_bbox(frame, boxes)
if inps is not None:
kps, kps_score, kps_id = self.pose_estimator.process_img(inps, boxes, pt1, pt2)
self.kps, self.kps_score = self.object_tracker.match_kps(kps_id, kps, kps_score)
return self.kps, self.id2bbox, self.kps_score
def classify_whole(self, pred_img, show_img):
pred = self.CNN_model.classify_whole(pred_img, show_img)
return pred
def classify(self, pred_img, show_img, id2bbox):
preds = self.CNN_model.classify(pred_img, id2bbox)
self.CNN_model.visualize(show_img, preds)
return preds
class HumanDetection:
def __init__(self, show_img=True):
self.object_detector = ObjectDetectionYolo(cfg=yolo_cfg, weight=yolo_weight)
self.object_tracker = ObjectTracker()
self.pose_estimator = PoseEstimator(pose_cfg=pose_cfg, pose_weight=pose_weight)
self.BBV = BBoxVisualizer()
self.KPV = KeyPointVisualizer()
self.IDV = IDVisualizer(with_bbox=False)
self.boxes = tensor([])
self.boxes_scores = tensor([])
self.img_black = np.array([])
self.frame = np.array([])
self.id2bbox = {}
self.kps = {}
self.kps_score = {}
self.show_img = show_img
def init_sort(self):
self.object_tracker.init_tracker()
def clear_res(self):
self.boxes = tensor([])
self.boxes_scores = tensor([])
self.frame = np.array([])
self.id2bbox = {}
self.kps = {}
self.kps_score = {}
def visualize(self):
img_black = cv2.imread('video/black.jpg')
if config.plot_bbox and self.boxes is not None:
self.frame = self.BBV.visualize(self.boxes, self.frame, self.boxes_scores)
# cv2.imshow("cropped", (torch_to_im(inps[0]) * 255))
if config.plot_kps and self.kps is not []:
self.frame = self.KPV.vis_ske(self.frame, self.kps, self.kps_score)
img_black = self.KPV.vis_ske_black(self.frame, self.kps, self.kps_score)
if config.plot_id and self.id2bbox is not None:
self.frame = self.IDV.plot_bbox_id(self.id2bbox, self.frame)
# frame = self.IDV.plot_skeleton_id(id2ske, copy.deepcopy(img))
return self.frame, img_black
def process_img(self, frame, gray=False):
self.clear_res()
self.frame = frame
with torch.no_grad():
if gray:
gray_img = gray3D(copy.deepcopy(frame))
box_res = self.object_detector.process(gray_img)
else:
box_res = self.object_detector.process(frame)
self.boxes, self.boxes_scores = self.object_detector.cut_box_score(box_res)
if box_res is not None:
self.id2bbox = self.object_tracker.track(box_res)
boxes = self.object_tracker.id_and_box(self.id2bbox)
inps, pt1, pt2 = crop_bbox(frame, boxes)
kps, kps_score, kps_id = self.pose_estimator.process_img(inps, boxes, pt1, pt2)
self.kps, self.kps_score = self.object_tracker.match_kps(kps_id, kps, kps_score)
return self.kps, self.id2bbox, self.kps_score